.

0 (hodnocen0 x )

(1) Půjčeno:1x 

BK

1st pub.

London : Sage Publications, 1999

viii, 266 s. : il. ; 24 cm

ISBN 0-7619-5890-8 (brož.)

Obsahuje bibliografii na s. 252-260 a rejstřík

000239900

Contents // Preface // IX // 1 Introduction 1 // 1.1 Multilevel analysis 1 // 1.1.1 Probability models 2 // 1.2 This book 3 // 1.2.1 Prerequisites 5 // 1.2.2 Notation 5 // 2 // Multilevel Theories, Multi-stage Sampling, // and Multilevel Models // 2.1 Dependence as a nuisance // 2.2 Dependence as an interesting phenomenon // 2.3 Macro-level, micro-level, and cross-level relations // 3 // 6 // 6 // 7 // 9 // Statistical Treatment of Clustered Data // 3.1 // 3.2 // 3.3 // 3.4 // 3.5 // 3.6 // Aggregation // Disaggregation // The intraclass correlation // 3.3.1 Within-group and between-group variance // 3.3.2 Testing for group differences // Design effects in two-stage samples // Reliability of aggregated variables // Within- and between-group relations // 3.6.1 Regressions // 3.6.2 Correlations // 13 // 13 // 15 // 16 // 18 // 21 // 22 // 24 // 26 // 27 // 31 // 3.6 .3 Estimation of within- and between-group correlations 33 // 3.7 // Combination of within-group evidence // 35 // 4 The Random Intercept Model 38 // 4.1 A regression model: fixed effects only 39 // 4.2 Variable intercepts: fixed or random parameters? 41 // 4.2.1 When to use random coefficient models? 43 // 4.3 Definition of the random intercept model 45 // 4.4 More explanatory variables 51 // 4.5 Within- and between-group regressions 52 // vi Contents // 4.6 Parameter estimation 56 // 4.7 �Estimating’ random group effects: posterior means 58 // 4.7.1 Posterior confidence intervals 60 // 4.8 Three-level

random intercept models 63 // 5 The Hierarchical Linear Model 67 // 5.1 Random slopes 67 // 5.1.1 Heteroscedasticity 68 // 5.1.2 Don’t force TOI to be 0! 69 // 5.1.3 Interpretation of random slope variances 70 // 5.2 Explanation of random intercepts and slopes 72 // 5.2.1 Cross-level interaction effects 73 // 5.2.2 A general formulation of fixed and random parts 79 // 5.3 Specification of random slope models 80 // 5.3.1 Centering variables with random slopes? 80 // 5.4 Estimation 82 // 5.5 Three and more levels 83 // 6 Testing and Model Specification 86 // 6.1 Tests for fixed parameters 86 // 6.1.1 Multi-parameter tests for fixed effects 88 // 6.2 Deviance tests 88 // 6.2.1 Halved p-values for variance parameters 90 // 6.3 Other tests for parameters in the random part 91 // 6.4 Model specification 91 // 6.4.1 Working upward from level one 94 // 6.4.2 Joint consideration of level-one and level-two variables 96 // 6.4.3 Concluding remarks about model specification 97 // 7 How Much Does the Model Explain? 99 // 7.1 Explained variance 99 // 7.1.1 Negative values of R2? 99 // 7.1.2 Definitions of proportions of explained variance // in two-level models 101 // 7.1.3 Explained variance in three-level models 104 // 7.1.4 Explained variance in models with random slopes 104 // 7.2 Components of variance 105 // 7.2.1 Random intercept models 106 // 7.2.2 Random slope models 108 // 8 Heteroscedasticity 110 // 8.1 Heteroscedasticity at level one 110 // 8.1.1 Linear variance functions

110 // 8.1.2 Quadratic variance functions 114 // 8.2 Heteroscedasticity at level two 119 // Contents V11 // 9 Assumptions of the Hierarchical Linear Model 120 // 9.1 Assumptions of the hierarchical linear model 120 // 9.2 Following the logic of the hierarchical linear model 121 // 9.2.1 Include contextual effects 122 // 9.2.2 Check whether variables have random effects 122 // 9.2.3 Explained variance 123 // 9.3 Specification of the fixed part 124 // 9.4 Specification of the random part 125 // 9.4.1 Testing for heteroscedasticity 126 // 9.4.2 What to do in case of heteroscedasticity 128 // 9.5 Inspection of level-one residuals 128 // 9.6 Residuals and influence at level two 132 // 9.6.1 Empirical Bayes residuals 132 // 9.6.2 Influence of level-two units 134 // 9.7 More general distributional assumptions 139 // 10 Designing Multilevel Studies 140 // 10.1 Some introductory notes on power 141 // 10.2 Estimating a population mean 142 // 10.3 Measurement of subjects 143 // 10.4 Estimating association between variables 144 // 10.4.1 Cross-level interaction effects 148 // 10.5 Exploring the variance structure 151 // 10.5.1 The intraclass correlation 151 // 10.5.2 Variance parameters 154 // 11 Crossed Random Coefficients 155 // 11.1 A two-level model with a crossed random factor 155 // 11.1.1 Random slopes of dummy variables 156 // 11.2 Crossed random effects in three-level models 159 // 11.3 Correlated random coefficients of crossed factors 160 // 11.3.1 Random slopes in a crossed design

160 // 11.3.2 Multiple roles 161 // 11.3.3 Social networks 162 // 12 Longitudinal Data 166 // 12.1 Fixed occasions 167 // 12.1.1 The compound symmetry model 168 // 12.1.2 Random slopes 171 // 12.1.3 The fully multivariate model 173 // 12.1.4 Multivariate regression analysis 178 // 12.1.5 Explained variance 179 // 12.2 Variable occasion designs 181 // 12.2.1 Populations of curves 181 // 12.2.2 Random functions 182 // 12.2.3 Explaining the functions 193 // viii Contents // 12.2.4 Changing covariates 195 // 12.3 Autocorrelated residuals 199 // 13 Multivariate Multilevel Models 200 // 13.1 The multivariate random intercept model 201 // 13.2 Multivariate random slope models 206 // 14 Discrete Dependent Variables 207 // 14.1 Hierarchical generalized linear models 207 // 14.2 Introduction to multilevel logistic regression 208 // 14.2.1 Heterogeneous proportions 208 // 14.2.2 The logit function: Log-odds 211 // 14.2.3 The empty model 213 // 14.2.4 The random intercept model 215 // 14.2.5 Estimation 218 // 14.2.6 Aggregation 219 // 14.2.7 Testing the random intercept 220 // 14.3 Further topics about multilevel logistic regression 220 // 14.3.1 Random slope model 220 // 14.3.2 Representation as a threshold model 223 // 14.3.3 Residual intraclass correlation coefficient 224 // 14.3.4 Explained variance 225 // 14.3.5 Consequences of adding effects to the model 227 // 14.3.6 Bibliographic remarks 229 // 14.4 Ordered categorical variables 229 // 14.5 Multilevel Poisson regression 234 // 15

Software 239 // 15.1 Special software for multilevel modeling 239 // 15.1.1 HLM 240 // 15.1.2 MLn / MLwiN 243 // 15.1.3 VARCL 245 // 15.1.4 MIXREG, MIXOR, MIXNO, MIXPREG 247 // 15.2 Modules in general purpose software packages 248 // 15.2.1 SAS, procedure MIXED 248 // 15.2.2 SPSS, command VARCOMP 249 // 15.2.3 BMDP-V modules 250 // 15.2.4 Stata 250 // 15.3 Other multilevel software 251 // 15.3.1 PinT 251 // 15.3.2 Mplus 251 // 15.3.3 MLA 251 // 15.3.4 BUGS 251 // References 252 // Index // 261